Projects
Stellantis Summer Intern Project
During my internship at Stellantis, I designed, programmed, and installed an AI vision based camera system for vehicle quality control. This project was designed to enhance the detection of issues during assembly by catching and addressing them during the assembly process rather than after the vehicle is built. Below is a step-by-step explanation of the work I conducted during the project.
Assess and Find Installation Location
The first step of the project involved evaluating potential locations for the camera installation. I spent the initial two weeks of my internship gaining an understanding of how the plant operated, observing the vehicle assembly process, understanding the role of quality control, and learning how to interpret the quality data produced by the plant’s control software. The primary objective was to identify a location that would allow us to capture and address as many potential issues as possible with a single installation.
Design Installation
After selecting the installation site, I began designing the frame for mounting the cameras. I created detailed engineering drawings for the contractors and a 3D-printed scaled-down model to demonstrate the design. The goal was to optimize the camera setup to monitor 75 fuel and electrical connectors using 15 cameras, while ensuring that the system would be durable enough for a factory environment. The design accounted for power needs, wiring, and accessibility for maintenance, and it was essential that the system did not interfere with existing equipment or operations.
Oversee Installation
Due to union regulations, we had to hire external contractors to handle the physical installation of the cameras. Once the design was finalized, I was responsible for overseeing the installation process. This involved coordinating with electricians, ensuring the installation followed design specifications, and troubleshooting any issues that arose during the process.
Linux Server Setup
I assisted with setting up and configuring the Linux server that hosted IBM’s AI software. This included installing necessary software, troubleshooting, and ensuring compatibility with the plant’s network and firewall rules. The server was critical to the system, as it processed the 15 camera feeds in real-time and published the results via MQTT.
Program AI Models
A significant part of the project involved programming the AI models that analyzed the visual data from the camera system. I focused on training and refining these models to accurately detect issues in the assembly process. This step required ongoing testing and optimization to ensure the system’s performance and reliability met expectations.
Conclusion
The project’s outcome resulted in an independent, locally controlled solution capable of improving quality control in the assembly process. This experience not only enhanced my technical skills but also gave me valuable insight into the complexities of industrial AI applications.
Circuit Board Business Cards
I wanted to come up with a way to stand out at career fair, I decided that the best way to do this would be with a super unique business card. I started with researching what people have done in the past, and I settled on either a nice aluminum etched card, or a circuit board, while aluminum would look nice, using a circuit board would allow my creativity to run wild. I wanted my card to be both functional and aesthetic, so I began by thinking about the different ways people could interact with it. I settled on three methods of interaction, first having my information written on the card in the conventional sense, second, having a QR code that creates a contact on your phone with my information pre-loaded, and finally, an NFC chip to link to this website. I designed the card using Autodesk Eagle. The card features an embedded loop antenna, meaning it can absorb electricity from your phone without making physical contact. When a smartphone is nearby, the card powers up, activates the NFC chip which sends my website and lights up the LED.
RAM 1500 Office Chair
One day during my summer internship at at Stellantis, I saw what looked like a perfectly good drivers seat sitting next to a dumpster with a note that said “broken” on it. I asked the repair tech what was wrong with the chair and he said there was a bad wiring harness. I then asked my manager if there was any way I could take home the “dumpster chair” to build an office chair, he said he didn’t have the ability to say yes or no, but he pointed me towards someone who could. I asked that person if I could take the chair out of the dumpster to turn into an office chair and he laughed, said yes, and did the appropriate paperwork.
Once I had the chair at home, I did some root cause analysis to figure out what was wrong with the wiring, turned out the front tilt motor wiring was completely missing, that was an easy 20 minute fix with a soldering iron and some spare wire I had. The next step was to reverse engineer the wiring harness, to try and get the motors, heating, and cooling to work. After several frustrating hours, I found the right wires, and soldered new connectors in place to turn on the heating and cooling. Unfortunately with the way the seat was designed, the seat movement switches went out to the cars computer, then the computer operated the motors, so I wasn’t able to use the original switches, but I added new switches and got all the motors rewired to my new switches. Finally, I attached the whole chair to an office chair base, and I have a fully functioning RAM 1500 seat as an office chair.
This was a fun project, because I got to take things apart to do root cause analysis and reverse engineering, I got to turn trash into something awesome, and there’s one less car part in a landfill.
Plowie
Who is Plowie? Plowie is an almost entirely 3D printed robot, whose primary objective is to be cute, easy to operate, and inexpensive to build. Purpose Plowie started as an offseason design project to hone my CAD skills, as I worked on it, it realized that it can also be an team mascot, ambassador, teaching tool, demonstration, and competition practice robot.
The inspiration for Plowie came from one of the animated robots in the First Tech Challenge game reveal animation. The animated robot has a square blue body, large front plow, and big cute eyes.
This robot was designed to be simple, reliable, and cheap. I decided to 3D print most of the robot to allow for design adaptability and because my team was sponsored by a 3D printing filament company. The front wheels are un-powered omni directional wheels, and the back wheels are each powered by one servo. The plow is powered by 2 servos, and the eyes are linked to one servo , which allows them to move. The robot has enough power to move itself around a room at a decent pace, but not enough power to do any damage. The control system is a REV expansion hub. I designed the robot so it can be operated with an Arduino if needed.
What did I learn? I learned a lot about designing a product for a specific manufacturing technique. In this case, I designed the robot around my 3D printer. I also designed the robot’s parts in ways that would allow them to be printed with minimal support material, resulting in a significant cost savings. I learned about how aesthetics can impact the functionality of a robot.
Infinity Cube
What is it and how does it work? Similar to an infinity mirror, the cube uses one-way mirrors facing inward to create the infinity effect. The cube is made of six pieces of one-way mirror, which allows you to see in from every side and allows the light to reflect back and forth inside infinitely in all three dimensions.
Keeping the price down was a big factor in the design of this project. I learned from my first mirror that one-way mirror film does not work as well as real one way glass, so I purchased real one-way mirror for this project. One-way mirrors are expensive, so I found a company that sold 6 inch squares for about 1/5th the cost of purchasing custom cut glass. I designed the frame in Autodesk inventor, and 3D printed it at home. Once I had all of the components, it took me about 10 hours to solder the 75 joints connecting the LEDs together. The LEDs I chose are digitally addressable, and plugged into an Arduino, running an open source project called WLED.
This project allowed me to test, and apply the skills I learned from the less successful infinity mirror project. I also continued to practice and improve my soldering skills.
Automated Wasabi Grow Pod
This was the project I completed for my ENG1101 class, the purpose of the project was to build an automated growing solution using sensors and an Arduino to control environmental factors, and increase plant growth efficiency. I designed an automated indoor growing system to cultivate real wasabi, a plant that requires very specific and stable conditions to thrive. Since most wasabi sold globally is an artificial blend of horseradish and food coloring, my goal was to make real wasabi more accessible by creating a controlled environment that mimics its natural habitat. I developed a modular system using a combination of off-the-shelf components and custom 3D-printed parts, with sensors to monitor water levels, temperature, humidity, light, and water quality. I programmed the control logic in MATLAB and ran it on an Arduino, which allowed for precise environmental control while optimizing energy efficiency. A key challenge was balancing the power consumption with the need to maintain ideal conditions, which I solved by completing a power budget and adding insulation around heating elements.
Throughout the project, I conducted a Failure Mode and Effects Analysis (FMEA) to identify potential risks, such as pump malfunctions or sensor failures, and implemented strategies to enhance system reliability. During testing I found several design oversights, and iterated on the original design. This project taught me valuable lessons in complex systems integration, risk management, and sustainable design while highlighting the potential to grow real wasabi in non-native regions. Future iterations will focus on further enhancing energy efficiency, reducing costs, and adding remote monitoring capabilities, ultimately working toward a viable commercial solution for indoor wasabi cultivation.
custom 3D Printer firmware
Inspiration
In 2020 I purchased an Ender 3 pro 3D printer to help my robotics team print face shields. Being the engineer I am, I was constantly tinkering with it, trying new things swapping out components, and trying new things. More specifically, I added better stepper motor drivers making the printer quieter, and an auto bed leveling sensor. The auto bed leveling sensor fundamentally changed how the printer calibrated itself before printing something. The manufacturer had a code for this, however after using it for a week, I wasn’t satisfied.
Many 3D printers run on an open-source software called Marlin. I took the stock Marlin project, modified it to work with my printer, and then wrote portions of code to implement the specific hardware I wanted to use. While I was working on it, I also made some modifications to how the user interface functions. One example is changing the preheating functions, making them easier to operate. I also added a fun boot screen which you can see in the photo to the left.
Infinity Mirror
This was one of my first larger scale projects. I found some videos online showing how to build and infinity mirror, a device that creates an illusion of infinite depth through the reflection of light between a standard mirror and a one-way mirror. I didn’t think it looked that hard, so my initial plan was to construct 10 smaller triangular infinity mirrors and assemble them into a single cohesive piece.
I quickly realized this project was way over my skillset, I encountered several challenges along the way. My initial plan to use acrylic proved problematic as mirror film did not adhere well enough to create the infinity effect. Soldering, which I had never done before, and assumed would be straightforward, turned out to be more complex. Additionally, cutting and drilling glass and aluminum presented unexpected difficulties. I ended up having to revise the design 3 times scaling down the complexity of the project. Despite the project’s final cost being approximately $900—far exceeding my initial goal of $250—I successfully completed the infinity mirror. While this project might not seem like much now, this was the first time I tried building something far outside my comfort zone, and taught me about the importance of failing to succeed. Through this project, I learned how to solder, cut glass, how different typed of adhesive work, and programming an Arduino, along with problem solving, iterative design, and troubleshooting.